[RESOLVED] PHP preg_match NOT equal to backreference

[tomcatexodus]

Is there a way to negate backreferences in a PHP preg_match test?

   .
   .
   .
1000| asdf34asdfasdf
1000| asdfadf43tasda34tfasdff
1000| qewrqwerqwt4er
1001| weklrjwelkr3jlkw
1001| welrjewrek455lwjrlkw
1001| trewewerwehklrjwelkrjlkw
1001| welrjewrek4hlwjrlkw
1001| fgfweklrjh4welkrjlkw
1001| welrjewh4reklwjrlkw
1002| weroiw5eewkjekwjw
1002| weroiweewkjekwjw
1002| we234roiweewkjekwjw
1002| weroiwe5ewkjekwjw
1002| weroiw324eewkjekwjw
   .
   .
   .

I want to be able to grab the data from the first instance of 1000 until the first instance of 1001, but using preg_match_all, continue to perform this for 1001, 1002, etc.

effectively producing an array containing:

[1]
1000| asdf34asdfasdf
1000| asdfadf43tasda34tfasdff
1000| qewrqwerqwt4er

[2]
1001| weklrjwelkr3jlkw
1001| welrjewrek455lwjrlkw
1001| trewewerwehklrjwelkrjlkw
1001| welrjewrek4hlwjrlkw
1001| fgfweklrjh4welkrjlkw
1001| welrjewh4reklwjrlkw

[3]
1002| weroiw5eewkjekwjw
1002| weroiweewkjekwjw
1002| we234roiweewkjekwjw
1002| weroiwe5ewkjekwjw
1002| weroiw324eewkjekwjw

Any ideas?

[kows]

and create this array with just a call to preg_match_all()? no. you'll need to process the array somehow:

<?php
$str = <<<EOF
1000| asdf34asdfasdf
1000| asdfadf43tasda34tfasdff
1000| qewrqwerqwt4er
1001| weklrjwelkr3jlkw
1001| welrjewrek455lwjrlkw
1001| trewewerwehklrjwelkrjlkw
1001| welrjewrek4hlwjrlkw
1001| fgfweklrjh4welkrjlkw
1001| welrjewh4reklwjrlkw
1002| weroiw5eewkjekwjw
1002| weroiweewkjekwjw
1002| we234roiweewkjekwjw
1002| weroiwe5ewkjekwjw
1002| weroiw324eewkjekwjw
EOF;
  
  $pattern = '/(\d{4})\| [a-z]+/i';
  
  preg_match_all($pattern, $str, $matches);
  
  // array and counter
  $parsed = array();
  $j = -1;
  $n = 0;
  
  // loop through
  for($i = 0; $i < count($matches[1]); $i++){
    
    $num   = $matches[1][$i];
    $match = $matches[0][$i];
    
    // if the number has changed:
    if($num != $n){
      
      // increase counter
      $j++;
      
      // keep a reference of the number we're on
      $n = $num;
      
      // "set" the initial state of this key
      $parsed[$j] = "";
    }
    
    // store entire match
    $parsed[$j] .= $match . "\n";
  
  }
  
  print_r($parsed);
  
?>

which produces:

Array
(
    [0] =>
1000| asdf
1000| asdfadf
1000| qewrqwerqwt

    [1] =>
1001| weklrjwelkr
1001| welrjewrek
1001| trewewerwehklrjwelkrjlkw
1001| welrjewrek
1001| fgfweklrjh
1001| welrjewh

    [2] =>
1002| weroiw
1002| weroiweewkjekwjw
1002| we
1002| weroiwe
1002| weroiw

)

[tomcatexodus]

Thanks kows (I should probably preface every thread with that )

So have you inadvertently answered the question of:

Is there a way to negate backreferences in a PHP preg_match test?

I was hoping that one could perform:

(?<id>[0-9]+?)\|(?<data>.*?)([^(P?<id>)])

Or something to that effect...

[tomcatexodus]

The BIG problem is... I'm working with files in excess of 70Mb...

The average number of lines per file is in excess of several hundred thousand.

I don't want the array index to overflow on me...

[kows]

you should give it a try and see if you do have an overflow then. you shouldn't, though. if you do, however, then you would have the same overflow if one regular expression did do it all anyway. the array would be the same size. if PHP is going to crash by you doing it manually, it's going to crash when it creates the array itself, too.